Position Interpolation Improves ALiBi Extrapolation

التفاصيل البيبلوغرافية
العنوان: Position Interpolation Improves ALiBi Extrapolation
المؤلفون: Al-Khateeb, Faisal, Dey, Nolan, Soboleva, Daria, Hestness, Joel
سنة النشر: 2023
المجموعة: Computer Science
مصطلحات موضوعية: Computer Science - Computation and Language, Computer Science - Artificial Intelligence, Computer Science - Machine Learning
الوصف: Linear position interpolation helps pre-trained models using rotary position embeddings (RoPE) to extrapolate to longer sequence lengths. We propose using linear position interpolation to extend the extrapolation range of models using Attention with Linear Biases (ALiBi). We find position interpolation significantly improves extrapolation capability on upstream language modelling and downstream summarization and retrieval tasks.
Comment: 4 pages content, 1 page references, 4 figures
نوع الوثيقة: Working Paper
URL الوصول: http://arxiv.org/abs/2310.13017
رقم الأكسشن: edsarx.2310.13017
قاعدة البيانات: arXiv